The underwater sound of R/V "Johan Hjort" as a function of aspect angle has been mapped by cruising along parallel transects at various distances from the recording unit. At frequencies over 125 Hz there was a minimum in front of the vessel, and lobes of higher intensity on both sides. About 20% of the herring schools recorded during a survey in the North Sea seemed to react to the survey vessel. The distance at which the schools reacted varied from about 25 m and up to about 1000 m in front of the vessel. Within a sector of 20° on each side of the path of the vessel, the fraction of schools reacting increased to about 50%. Most of the schools that reacted seem to be herded in front of the vessel. This reaction pattern may indicate either that the schools reacted to the higher sound intensity to the side of the vessel or that the pattern occurred because of a "pursuit effect". Despite the apparent herding, more than 40% of the schools in front of the vessel were not recorded by the echo sounder. It is therefore argued that horizontal avoidance close to the vessel may have caused an underestimation of the biomass of herring of about 20%.